Opened 3 years ago

Closed 18 months ago

#1857 closed Bug (fixed)

ssh=0 in the corners with BDY

Reported by: clem Owned by: gmattia
Priority: normal Milestone:
Component: OCE Version: release-3.6
Severity: minor Keywords:
Cc: gmattia

Description

Context

I run a regional configuration with 4 boundaries using BDY (both in 3.6 stable and in the trunk) and I encounter a problem in the 4 corners of my domain: SSH=0 there while it should not be.

Analysis

The issue only occurs with BDY (a closed boundary simulation is OK)
It is not due to the tides (removed).
I checked the input files which are OK at the corners (SSH/=0).

Fix

Does anybody know where it can come from?

Commit History (0)

(No commits)

Change History (7)

comment:1 Changed 3 years ago by clem

In addition, I do not understand in the trunk the complexity of the subroutine bdy_ssh (in bdydyn2d.F90).

The following lines:

            zcoef1 = bdytmask(ii-1,ij  ) +  bdytmask(ii+1,ij  )
            zcoef2 = bdytmask(ii  ,ij-1) +  bdytmask(ii  ,ij+1)
            IF ( zcoef1+zcoef2 == 0 ) THEN
               ! corner
               zcoef = bdytmask(ii-1,ij) + bdytmask(ii+1,ij) +  bdytmask(ii,ij-1) +  bdytmask(ii,ij+1)
               zssh(ii,ij) = zssh(ii-1,ij  ) * bdytmask(ii-1,ij  ) + &
                 &           zssh(ii+1,ij  ) * bdytmask(ii+1,ij  ) + &
                 &           zssh(ii  ,ij-1) * bdytmask(ii  ,ij-1) + &
                 &           zssh(ii  ,ij+1) * bdytmask(ii  ,ij+1)
               zssh(ii,ij) = ( zssh(ii,ij) / MAX( 1, zcoef) ) * tmask(ii,ij,1)

Can be replaced by:

            zcoef = bdytmask(ii-1,ij) + bdytmask(ii+1,ij) +  bdytmask(ii,ij-1) +  bdytmask(ii,ij+1)
            IF ( zcoef == 0 ) THEN
               zssh(ii,ij) = 0._wp

Or do I miss something here?

Last edited 3 years ago by nicolasmartin (previous) (diff)

comment:2 Changed 3 years ago by clevy

  • Owner changed from nemo to gmattia

comment:3 Changed 3 years ago by clevy

  • Cc gmattia added
  • Status changed from new to assigned

comment:4 Changed 3 years ago by gmattia

I’d suggest to simplify the subroutine bdy_ssh (bdydyn2d.F90 in the trunk) as it is done in 3.6 stable version.
In the 4 corners of domain SSH=0, it's no correct.
In subroutine bdy_ssh (in 3.6 stable version ), when zcoeff=0 zssh(ii,jj)=0, but I don't understand.

comment:5 Changed 3 years ago by gmattia

  • Owner changed from gmattia to clevy

In subroutine bdy_ssh (in 3.6 stable version ), when zcoeff=0 zssh(ii,jj)=0,
but I still do not understand the motivation.

comment:6 Changed 3 years ago by clevy

  • Owner changed from clevy to gmattia

The fix is now described in ticket 1997 and should be applied both on 3_6_STABLE, in the trunk and in devmerge_2017 branch

comment:7 Changed 18 months ago by clem

  • Resolution set to fixed
  • Status changed from assigned to closed

fixed at r10518

Note: See TracTickets for help on using tickets.